VERtical Innovation in the Domain of Robotics Enabled by Artificial intelligence Methods

Deliverables

A report on the scope and limitation of the DREAM methods along with pragmatic guidelines for deploying and adapting them to different types of use cases

Scientific paper or report on generalisation and robustification through exploration

A collection of scientific papers on individual methods or an overview article over all methods The deliverable is relatively late in the project to account for reviewing periods

Broad Exploitation Strategy Report

Report describing the strategy for exploiting the innovation potential of VERIDREAM outside the boundaries of the projectAfter Synesis participation in the project has ended the content of this deliverable has been refined to include Interviews with academic leaders who have made a startup in AI and their lessons learned from that eg Pieter Abbeel George Konidaris A strategy for patenting AI which is not possible in the EU as such when applied to robotics which facilitates patenting AI A concept for IP management wrt to academic opensource software as background in EU project and its compatibility with DESCA A metaanalysis of the innovation loop ie the main lessons learned from it

Scientific paper or report on generalisation and robustification through representation

A collection of scientific papers on individual methods or an overview article over all methods The deliverable is relatively late in the project to account for reviewing periods

Coordination with the AI4EU platform

Document describing the interaction with the AI4EU platform in terms of dissemination as well as the strategy for publishing data and software on AI4EU

Scientific workshop

A scientific workshop to be organised at IROS ICRA RSS or ICDLEpiRob

Project Website Publication

Document describing the interaction with the AI4EU platform in terms of dissemination as well as the strategy for publishing data and software on AI4EU

Demonstration of GoodAI Use Cases

A demonstration of the DREAM methods within the Space Engineers video game.

Publications

Stable-Baselines3: Reliable Reinforcement Learning Implementations

Author(s): Raffin, Antonin; Hill, Ashley; Gleave, Adam; Kanervisto, Anssi; Ernestus, Maximilian; Dormann, Noah
Published in: Journal of Machine Learning Research (JMLR), Issue 6, 2021, ISSN 1533-7928
Publisher: Microtome Publishing

Few-shot Quality-Diversity Optimization

Author(s): Achkan, Salehi; Salehi, Achkan; Coninx, Alexandre; Doncieux, Stephane
Published in: https://hal.archives-ouvertes.fr/hal-03569179, Issue 3, 2022, ISSN 2377-3766
Publisher: Institute of Electrical and Electronics Engineers Inc.
DOI: 10.1109/lra.2022.3148438

Smooth Exploration for Robotic Reinforcement Learning

Author(s): Raffin, Antonin; Kober, Jens; Stulp, Freek
Published in: Conference on Robot Learning, Issue 3, 2022, Page(s) 1634-1644, ISSN 2640-3498
Publisher: PMLR
DOI: 10.48550/arxiv.2005.05719

Innovation Paths for Machine Learning in Robotics

Author(s): Stulp, Freek; Spranger, Michael; Listmann, Kim; Doncieux, Stéphane
Published in: IEEE Robotics & Automation Magazine, Issue 29(4), 2022, Page(s) 141--144, ISSN 1070-9932
Publisher: Institute of Electrical and Electronics Engineers
DOI: 10.1109/mra.2022.3213205

Fault-Tolerant Six-DoF Pose Estimation for Tendon-Driven Continuum Mechanisms

Author(s): Antonin Raffin, Bastian Deutschmann, Freek Stulp
Published in: Frontiers in Robotics and AI, Issue 8, 2021, ISSN 2296-9144
Publisher: Frontiers Media
DOI: 10.3389/frobt.2021.619238

Exploratory State Representation Learning

Author(s): Astrid Merckling; Nicolas Perrin-Gilbert; Alex Coninx; Stéphane Doncieux
Published in: Frontiers in Robotics and AI, Vol 9 (2022), Issue 8, 2022, ISSN 2296-9144
Publisher: Frontiers Media SA
DOI: 10.48550/arxiv.2109.13596

Learning to Exploit Elastic Actuators for Quadruped Locomotion

Author(s): Antonin Raffin, Daniel Seidel, Jens Kober, Alin Albu-Schäffer, João Silvério, Freek Stulp
Published in: 2022
Publisher: arxiv
DOI: 10.48550/arxiv.2209.07171

Apprendre aux robots à faire face à l'imprévu

Author(s): Stephane Doncieux
Published in: Industries et Technologies [Cahier Technique], Issue 1045, 2021
Publisher: Infopro

Transferring AI research results into an industrial product

Author(s): "Steffen R\""uhl, Moritz Tenorth, Achkan Salehi, Stéphane Doncieux"
Published in: White paper, 2023
Publisher: Magazino

Adaptive Asynchronous Control Using Meta-learned Neural Ordinary Differential Equations

Author(s): Achkan Salehi, Steffen Rühl, Stéphane Doncieux
Published in: 2022
Publisher: arxiv

Reinforcement Learning Guided by Shared Control Templates

Author(s): Abhishek Padalkar, Gabriel Quere, Franz Steinmetz, Antonin Raffin, Matthias Nieuwenhuisen, Joao Silverio, Freek Stulp
Published in: Proceedings of the IEEE-ICRA conference, 2023
Publisher: IEEE

BR-NS - an archive-less approach to novelty search

Author(s): Achkan Salehi, Alexandre Coninx, Stephane Doncieux
Published in: Proceedings of the Genetic and Evolutionary Computation Conference, 2021, Page(s) 172-179, ISBN 9781450383509
Publisher: ACM
DOI: 10.1145/3449639.3459303

Closing the Gap: Combining Task Specification and Reinforcement Learning for Compliant Vegetable Cutting

Author(s): Padalkar, Abhishek; Nieuwenhuisen, Matthias; Schulz, Dirk; Stulp, Freek
Published in: Informatics in Control, Automation and Robotics ISBN: 9783030924416, Issue 3, 2022, ISBN 9783030924416
Publisher: Springer
DOI: 10.1007/978-3-030-92442-3_11

Smooth Exploration for Robotic Reinforcement Learning

Author(s): Antonin Raffin and Jens Kober and Freek Stulp
Published in: Conference on Robot Learning, 2021
Publisher: openreview.net

Deliverables

Publications

Share this page

Download